Download Fiscal Consolidation Strategy

Fiscal Consolidation Strategy
John F. Cogan, John B. Taylor, Volker Wieland, Maik Wolters*
June 14, 2012
Abstract
In the aftermath of the global financial crisis and great recession, many countries face
substantial deficits and growing debts. In the United States, federal government outlays as a
ratio to GDP rose substantially from about 19.5 percent before the crisis to over 24 percent
after the crisis. In this paper we consider a fiscal consolidation strategy that brings the budget
to balance by gradually reducing this spending ratio over time to the level that prevailed prior
to the crisis. A crucial issue is the impact of such a consolidation strategy on the economy.
We use structural macroeconomic models to estimate this impact. We consider two types of
dynamic stochastic general equilibrium models: a neoclassical growth model and more
complicated models with price and wage rigidities and adjustment costs. We separate out the
impact of reductions in government purchases and transfers, and we allow for a reduction in
both distortionary taxes and government debt relative to the baseline of no consolidation.
According to the initial model simulations GDP rises in the short run upon announcement and
implementation of this fiscal consolidation strategy and remains higher than the baseline in
the long run.
* John F. Cogan is the Leonard and Shirley Ely Senior Fellow at the Hoover Institution and a
professor in the Public Policy Program at Stanford University. John B. Taylor is the George P. Shultz
Senior Fellow in Economics at the Hoover Institution Mary and Robert Raymond Professor of
Economics at Stanford University. Volker Wieland is Professor of Monetary Economics at the
Institute for Monetary and Financial Stability (IMFS) at Goethe University of Frankfurt. Maik Wolters
is a Post-Doctoral Researcher at the IMFS. Helpful comments by Günter Coenen, Jan in ’t Veld,
Roland Straub, Harald Uhlig and participants of the Fiscal Policy in the Aftermath of the Financial
Crisis Conference at the European Commission in Brussels, March 2-3, 2012 are gratefully
acknowledged. All remaining errors are the authors’ sole responsibility.
1
As a consequence of the financial crisis and great recession government deficits have
risen substantially creating the need for a fiscal consolidation strategy to reduce the deficits
and stop the growing debt. This increase in budget deficits resulted partly from greater
spending and transfers and partly from lower tax receipts during the recession. Looking
forward, sustained spending increases are particularly worrisome, because they ultimately
require raising tax rates beyond pre-crisis levels, even after the economic recovery. Higher
distortionary taxes may then dampen the economy’s trend growth for a long time.
Figure 1 summarizes the federal government spending situation in the United States. It
shows U.S. federal government outlays—excluding interest payments—relative to GDP.
Government outlays (or government spending) include both government transfers and
government purchases of goods and services. The history line shows the rapid increase in
spending during the crisis. The baseline shows that the high spending levels are projected to
continue unless current federal spending policies are changed.1 The interest payments on
federal debt that is expected to be issued to finance this high level of spending, (not shown in
the chart) will add increasingly larger amounts to total federal spending. Under the baseline
assumptions, interest payments rise from 1.4 percent of GDP in 2012 to 3.7 percent in 2022.
Figure 1. U.S. Federal Outlays as a Percent of GDP
(excluding interest)
Percent
24
23
22
Baseline
21
20
History
19
Fiscal Consolidation
Strategy
18
17
16
15
00
1
02
04
06
08
10
12
Year
14
16
18
20
The details about the baseline and the consolidation strategy are presented below.
2
22
Implicit in this baseline is a long-run increase in tax rates needed to reduce the deficit and
thereby prevent the national debt from growing to economically dangerous levels. However,
higher tax rates themselves will distort private incentives for saving, investment and capital
accumulation to the detriment of economic growth and welfare.
Figure 1 also shows the path of federal government spending under a fiscal
consolidation strategy that gradually returns spending to the pre-financial crisis level as a
share of GDP. Because the U.S. federal budget was close to balance before the crisis, (the
federal deficit was only 1.3 percent of gdp in 2007) this strategy would mitigate the size of
any tax rate increase. Hence, relative to the current policy baseline, long-run tax rates would
be lower under this alternative strategy.
The purpose of this paper is to evaluate the impact of a fiscal consolidation plan on the
U.S. economy, including quantifying its impact on GDP, consumption and investment. Of
course, the size and sign of this impact is a crucial and widely debated policy question, which
is at the heart of the current austerity debate. We use modern structural macroeconomic
models to assess the impact.
To help establish the robustness of our results, we consider different types of models.
First we consider a neoclassical growth model with flexible prices and forward-looking
households. Such a model is helpful for clarifying some of the longer-run implications of
changes in government spending and tax cuts. For example, in this model, a reduction in
future government spending allows for a decrease in tax rates which will increase
employment and GDP; if that decrease in taxes is anticipated, then consumption will increase
in the short run.
Second, we consider forward looking models that include various price and wage
rigidities and adjustment costs not present in the neoclassical growth model. Such models are
needed to assess the short run impacts of a fiscal consolidation strategy. We focus on the
3
model of Coenen, McAdam and Straub (2008) and also consider the model of Cogan, Cwik,
Taylor and Wieland (2010) (CCTW).
The model of Coenen et al (2008) accounts for a range of distortionary taxes. It covers
the U.S. economy as well as the euro area economy as a whole, and was used by its European
Central Bank authors to investigate the impact of a reduction in distortionary taxes in the euro
area. It is sometimes called the New-Area-Wide Model (NAWM) since a version of the
model has been estimated with euro zone data and has replaced the so-called Area-WideModel (AWM) in policy.2 For the purpose of this study, we calibrate the parameters of the
U.S. part of the model with references to empirically estimated models, including the CCTW
model. The CCTW model is very similar to the well-known models of Christiano,
Eichenbaum and Evans (2005) and Smets and Wouters (2007). We also use simulations of the
CCTW model as a robustness check for the following two reasons. It is estimated by Bayesian
methods on the same U.S. data as in Smets and Wouters (2007). Furthermore, this model
includes not only forward looking households but also households that choose to consume
their current income. Due to the presence of these Keynesian or “rule-of-thumb” households
the CCTW model does not exhibit Ricardian equivalence. .
The paper is organized as follows. First, we present the neoclassical model, illustrate
its properties with some simple examples, and explore the sensitivity of these properties to
assumptions about the household utility function. Second, we present and illustrate the
properties of the models with certain rigidities. Third, we present the fiscal consolidation
strategy and compare it with the baseline assumption of what would happen if there were no
consolidation. Finally, we examine the effect of the strategy on the economy using the
models.
2
We have made our implementations of the AWM, NAWM and CCTW models available online in a new
macroeconomic model archive (see http://macromodelbase.com.). The model comparison approach is presented
in Taylor and Wieland (2012) and Wieland et al (2012).
4
1. The Neoclassical Model
This model builds on the model in King, Plosser and Rebelo (1988) and the discussion
in the Ljungvist and Sargent (2004) textbook. The government in this model buys goods and
finances these purchases with distortionary taxes on consumption, capital and labor, as well as
some non-distortionary lump-sum per-capita taxes. Government consumption is denoted by
g t and lump-sum taxes by  ht .  ct is the consumption tax rate,  kt the capital tax rate and  lt
the labor tax rate.
Households have preferences over consumption and leisure:

  U (c , l ),
t
t 0
t
t
  (0,1)
(1)
where ct denotes private consumption and lt hours worked. For the period utility function we
use the following standard specification:
U (c t , l t ) 
ct1 1
l 1 2

1 1 1  2
U (ct , l t )  ln ct  
l 1 2
1  2
for 0   1  1 and  1  1
for  1  1 .
(2)
.
 1 denotes the intertemporal elasticity of substitution, while  2 refers to the inverse of the
labor supply elasticity with respect to the real wage.3 The labor supply elasticity plays an
important role regarding the impact of changes in government spending on work hours and
total output, as we discuss in detail below.
The household’s budget constraint is given by

 p (1  
t 0
t

ct
)ct  pt it    rt (1   kt )k t  wt (1   lt )l t  pt ht ,
(3)
t 0
while the government has to satisfy
3
Another popular utility specification is given by U ( c t , l t )  ( 1   1 )  1 c t1     ( 1  l ) . Our
functional form includes the inverse of the labor supply elasticity directly as a parameter. It is more
convenient for investigating the implications of different labor supply elasticities.
1
5


t 0
t 0
 pt gt   ct pt ct  rt kt kt  wt lt lt  pt ht .
(4)
Here, p t denotes the time period t pre-tax price of one unit of investment, consumption or
government spending. wt refers to nominal pre-tax wages and rt to the nominal pre-tax rental
rate of capital. The consumption, capital and labor tax rates are set exogenously. Lump-sum
taxes are a residual used to balance the government budget given the exogenously specified
paths for government spending and distortionary tax rates ( g t , ct , kt , lt ).
Output is produced with a standard Cobb-Douglas production technology:
y t  F ( k t , l t )  k t l t(1 ) .
(5)
Capital is accumulated according to the following equation,
k t 1  (1   )k t  it ,
(6)
where k t denotes the capital stock, it investment and  the depreciation rate of capital. Market
clearing requires that
y t  c t  it  g t .
(7)
In equilibrium a representative household chooses ct , lt , it t 0 to maximize utility

defined by equations (1) and (2) subject to equations (3) and (6). A representative firm
chooses k t , lt t 0 to maximize profits,


 p y
t 0
t
t
 rt k t  wt l t  , subject to the production
function, equation (5). A feasible government policy is an expenditure and tax plan
g t , ct , kt , lt , ht t0 that satisfies its budget constraint, equation (4). A feasible allocation is a
sequence ct , it , yt t 0 that satisfies market clearing, i.e. equation (7).

6
In our initial analysis we abstract from uncertainty. The equilibrium outcome can be
characterized by the following conditions. First, consumption decisions must satisfy the
standard Euler equation,
ct1  ct11 Rt 1 ,
(8)
where Rt is the after-tax one-period gross interest rate between t and t + 1 measured in units
of consumption goods at t + 1 per consumption good at t:
 1
 kt 1  
(1   ct ) 
 .
(1   )  (1   kt1 ) 
Rt 
lt 1  
(1   ct1 ) 



(9)
Secondly, the consumption-leisure choice is determined by equating the marginal rate of
substitution with the real after-tax wage adjusted for consumption tax payments. In the
resulting equation (11), we have replaced the real wage by the marginal product of labor:
k
l t 2
(1   lt )
(1   ) t

 1
(1   ct )
ct
 lt


 .

(10)
Thirdly, capital evolves according to the following law of motion:
k t 1  k t l t1  (1   ) k t  ct  g t
(11)
Other equations are not necessary to obtain a solution to the model, but they are used to
determine the paths of other variables. Output can be obtained from equation (5) and
investment from equation (6). The real pre-tax rental rate of capital is given by the marginal
product of capital. The real pre-tax wage corresponds to the marginal product of labor.
The parameter values for the discount factor, the depreciation rate and the capital share
are set to the same values as in Chapter 11 of Ljungvist and Sargent (2004):
  0.95,   0.2,   0.33 . Regarding the utility parameter we set the intertemporal elasticity
of substitution  1  1 which corresponds to log utility. Log utility is compatible with a
7
balanced growth path (see King, Plosser and Rebelo, 1988).4 For the parameter governing the
labor-leisure decision, we use Smets and Wouters (2003, 2007) prior estimate of  2  2 . This
value implies a Frisch labor supply elasticity of 1 /  2  0.5 , which is consistent with
microeconomic estimates (see Chetty et al (2010)).
We calibrate the tax rates so that they match current U.S. tax rates. Specifically, we
use the U.S. values from the model of Coenen et al. (2008). Thus, the consumption tax rate is
7.7%, the labour tax rate is 22.5% and the capital tax rate is 18.41%.5 Other taxes or transfers
are collected lump-sum.
For illustrative purposes we consider a lasting reduction in government spending of 1
percent of GDP that is phased in very gradually over five years. Thus, during the initial 20
quarters government spending is reduced by 0.05% per quarter. This policy is announced in
the first period and anticipated by market participants from then onwards. Figure 2 shows the
impact on government spending, consumption, investment and total output. Percentage
changes in government spending, consumption and investment are weighted by the shares of
the respective variables in initial steady-state GDP. Thus, their values sum up to the
percentage change in total output relative to the initial steady-state.
In this example simulation, the reduction in government spending is assumed to be
due to a reduction in transfers. The funds from this reduction in transfers are used to reduce
other taxes. Here, we use them to lower the labor income tax which distorts household
decisions and induces adverse effects on the economy. Lowering such taxes may thus raise
economic production and welfare. Tax rates in our parameterization reflect current U.S. tax
4
We have also run all simulations with an inter-temporal elasticity of substitution,  1  2 . The results
are very similar and the differences are quantitatively small. Therefore, we focus on the log utility
case.
5
Coenen et al match the consumption tax, labor tax and social security contributions to data on U.S.
tax rates and euro area tax rates. They set the capital tax rate to the same value in both countries and
determined by matching the investment/output ratio.
8
rates and are drawn from Coenen et al (2008).6 Labor tax rates shown in Figure 2 are net of
social security contributions.
Figure 2: A reduction in transfers with savings applied to labor taxes
-1% of GDP, phased-in over 5 years
Consumption
Transfers
0
-0.5
-1
0
Investment
1
1
0.5
0.5
0
20
40
0
0
Output
20
40
0
Outlays/GDP
40
Labor Tax Rate
0
1
20
0.155
0.15
-0.5
0.5
0.145
0
0.14
-1
0
20
40
0
20
40
0
20
40
The reduction in transfers combined with a policy of using the budget savings to
finance a lower distortionary labor income tax induces a substantial lasting positive effect on
consumption, investment and total output. Households supply more labor. Higher investment
leads to greater capital stock. Government purchases remain constant, but overall government
outlays decline as a share of GDP by about 1 percentage point.
There are alternative tax and spending assumptions to go along with this illustrative
scenario. If government purchases are reduced rather than transfers the positive impact of on
GDP is smaller. If lower funding needs for purchases are channeled to reduce lump sum
rather than labor taxes, total GDP even falls in this model. This occurs because there is no
6
See Coenen et al (2008) for an evaluation of the effects of a reduction of Euro area tax rates to U.S. levels. All
tax rates reflect current US taxes except for the capital tax rate. The capital tax rate is determined by the
calibration steady state ratios of the model. Furthermore, a recent paper byUhlig and Drautzburg (2010)
investigates the negative impact from (future) increases in distortionary taxes needed finance fiscal stimulus
packages such as the ARRA legislation from 2009.
9
reduction in distortions and because the positive private sector wealth increase (due to the
decline in government spending) induces a decline in labor supply. Alternatively capital tax
rates can be reduced. Figure 3 shows the impact of a decline in the capital tax rate phased-in
over five years along with the decline in government spending, which, in this case, takes the
form of a decline in purchases rather than transfers. We consider reductions of 1 and 2
percentage points, respectively, down from an initial rate of 18.4%. As the savings from
government spending cuts are used to lower capital taxation, we observe an increase in the
capital stock. The depreciation rate is constant and thus in the new steady state more
investment is needed to keep the capital stock at the new, higher level. A decrease of the
capital tax rate of 2 percentage points leads to an increase in output.
Figure 3: Reducing government purchases and capital taxation
g reduced by 1% of GDP, capital tax by 1 and
2 percentage points respectively, phased-in over 5 years
Capital Tax Rate
Government spending
Capital Stock
0.19
0
3
0%
1%
2%
0.18
-0.5
2
1
0.17
0
-1
0
20
40
0.16
0
Consumption
20
40
-1
0
Investment
0.5
0.5
0.5
0
0
0
-0.5
-0.5
20
40
0
20
40
Output
1
0
20
40
0
20
40
Implications of a weaker labor supply elasticity
Note in the lower right hand graph in Figure 3 that output actually declines when the
budget savings are used to reduce lump sum taxes, instead of the distortionary capital tax rate.
This result occurs because, with government purchases not in the utility function, household
10
are wealthier when government purchases decline and transfer payment rise. Thus they
reduce their labor supply due to an income effect.
In the benchmark calibration we have set  2 equal to 2, which implies a labor supply
elasticity of 0.5. While this is a widely-used value in models (see e.g. Smets and Wouters
(2007) used to study macroeconomic fluctuations, 7 it might be too high to capture labor
supply responses over coming years when the economy may require substantial shifts in
sectoral labor allocation following the financial crisis and the great recession. Jobholders may
be less willing to reduce hours or quit their job compared to earlier decades, and there may be
larger inflow of new entrants that will render total work hours rather inelastic relative to a
reduction in the size of government.
To illustrate the implications of different assumptions about labor supply elasticity we
simulated versions of the model with inflexible labor supply, totally elastic labor supply and
intermediate cases of labor market responses, Thus, we compare values of  2  2 ,  2  4
(low elasticity of 1 /  2  0.25 ) and  2   (totally inelastic, 1 /  2  0 ).
Figure 4 shows the simulation results of a decline in government purchases with two
cases of a less elastic labor supply. A comparison with Figure 3 shows that consumption and
output increase by a larger amount when labor supply is less elastic.8
7
King and Rebelo (1999) set the labor supply elasticity to 4 and Cho and Cooley (1994) to 2.61. Chetty et al.
(2011 a,b) review the micro and macro evidence of labor supply elasticities. The micro evidence points to much
lower numbers, around 0.5 at the intensive margin and 0.25 at the extensive margin which add up to ¾ for
aggregate hours. Heterogeneity of elasticities between different groups of workers lead to the divergence of
micro and macro estimates. Extensive margin elasticities of prime-age men are close to zero.
8
The resulting tax revenues are shown in the Appendix.
11
Figure 4: Lower labor supply elasticity
g reduced by 1% of GDP, capital tax by 2 percentage points, phased-in over 5 years
Capital Tax Rate
Government spending
-0.5
Capital Stock
0.19
0
elasticity=0.5
elasticity=0.25
totally inelastic
5
4
0.18
3
2
0.17
-1
0
1
0
0.16
40
0
20
Consumption
20
40
0
Investment
1.5
1.5
1
1
1
0.5
0.5
0.5
0
0
0
20
40
0
20
40
Output
1.5
0
20
40
0
20
40
Implications of adding government consumption to household utility
The model can also be expanded to let household utility depend positively on
government consumption. If households derive utility from government consumption such as
infrastructure, police, fire protection, national defence, education they may respond to a
government spending cut with more private consumption to make up for it. We extend the
simple neoclassical model by introducing a new utility function which comprises a bundle of
private and public consumption c~t 9:

  U (c~ , l ),
t
t 0
t
t
  (0,1)
9
(12)
Earlier research considering government consumption in household utility includes Barro (1981), Barro (1990),
Christiano and Eichenbaum (1992), Baxter and King (1993), Ambler and Paquet (1996) and Finn (1998). These
studies treat government consumptions as perfect substitutes. Public spending categories like free school lunches
are close substitutes to private spending, while others like spending on transportation are probably complements
(see Karras, 1994). Kormendi (1983) and Aschauer (1995) find evidence for a substantial amount of
substitutability of private and public consumption for the US and Ahmed (1986) for the UK. Karras (1994)
examines evidence from a number of countries and finds that private and government consumption are best
described as complementary or unrelated goods. Ni (1995) finds evidence for complementarity between private
and public consumption. Amano and Wirjanto (1998) find that additive separability, i.e. public and private
consumption are unrelated, cannot be rejected.
12
Specifically, we can follow Ni (1995), Amano and Wirjanto (1998) and Linnemann
and Schabert (2004) and use a linearly homogenous consumption bundle of the CES form that
allows private and public consumption to be perfect or imperfect substitutes:

c~t  c~t (ct , g t )  ct  (1   ) g t

1/ 
,   ( ,1),   (0,1) .
(13)
 denotes the relative weight of private to public consumption.  determines the intratemporal elasticity of substitution   1 /(1   )  0 between private and public consumption.
  1 implies perfect substitutes.10 With   1 government consumption drops out of the
consumption bundle and utility simplifies to the standard case with private consumption and
leisure only. In terms of parameterization, we follow Amano and Wirjanto (1998) and use
their estimate of   0.36 . We then conduct a sensitivity study for different weights  of
private relative to public consumption in the utility function.
Figure 5 shows the effects of a joint reduction of government purchases and the
capital tax rate in the case where  =0.75 . The labor supply elasticity is set to 0.5 as in the
benchmark simulation. The simulation indicates that consumption and investment rise by a
larger amount when households derive utility from government consumption.11 The steadystate capital stock and output also increase.
10
Amano and Wirjanto (1998) show that the sign of the partial cross-derivative
U cg ,t  U (ct , g t ) / ct  / g t is determined by the relative magnitude of the intertemporal and intraperiod
elasticities of substitution: if 1 /  1   then private and public goods are complements, i.e. U cg ,t  0 , if
1 /  1   , i.e. U cg ,t  0 , then the two goods are substitutes and if 1 /  1   , i.e. U cg ,t  0 , then the goods
are unrelated.
11
Public and private consumption are substitutes (compare also appendix 2) and thus a decrease of utility due to a
decrease in public consumption can be compensated through an increase in private consumption. For private
consumption to increase sufficiently, produced output must not fall too much or even increase. This leads to an
increase in the capital stock and investment. Figure A4 in the appendix shows that the reduction in lump-sum tax
revenues is smaller the larger the value ofα.
13
Figure 5: Government purchases and tax cuts when households
derive utility from government purchases
g reduced by 1% of GDP, capital tax by 2 percentage points, phased-in over 5 years
Capital Tax Rate
Government spending
Capital Stock
0.19
0
alpha=1
alpha=0.75
5
4
0.18
3
-0.5
2
0.17
-1
0
20
40
0.16
0
Consumption
1
0
20
40
0
Investment
1.5
1.5
1
1
1
0.5
0.5
0.5
0
0
0
20
40
0
20
40
Output
1.5
0
20
40
0
20
40
2. Models with Rigidities
While the simple neoclassical growth model used in the preceding section is helpful for
illustrating some of the long-run consequences of changes in the fiscal policy regime, the
assumptions of perfect markets and flexible prices are not appropriate for assessing the short
run impacts. Thus, we consider models that include various price or wage rigidities and
adjustment costs.
The NAWM model of Coenen et al (2008) accounts for a range of distortionary taxes. As
with the neoclassical model, households maximize utility over time. But only some
households have access to financial markets; they can accumulate capital and hold money.
The other type of households can smooth consumption only by adjusting their money
holdings. Still, these households are more sophisticated than the rule-of-thumb households in
the CCTW model. The utility function also allows for slow adjustment of consumption due to
habits in which consumption at different periods of time affects current utility.
14
Households supply differentiated labor services and have some market power in wage
setting. Wages are determined in staggered contracts setting of the Calvo (1983) variety in
which some workers have an opportunity to change their wage while others have their wages
indexed to a geometric average of past changes in the prices. Households’ gross income is
subject to a variety of taxes. Households pay taxes on consumption purchases, on wage
income and on capital income. Furthermore, they pay social security contributions, a lumpsum tax and receive transfers. Purchases of consumption goods, financial investment in
international markets and capital utilization are subject to specific proportional adjustment
costs.
Firms produce tradable or non-tradable goods. Intermediate goods firms produce a
single, tradable differentiated good using an increasing-returns-to-scale Cobb-Douglas
technology with capital services and labor as inputs. These goods are sold in domestic and
foreign market under monopolistic competition. Like wages, price setting is subject to
staggered price contracts of the Calvo (1983) variety. Some firms get to adjust their price
each period, while other firms’ prices are indexed to a geometric average of past changes in
the aggregate price indexes. The final goods firms produce three non-tradable final goods:
private consumption goods, investment goods and public consumption goods. Final nontradable private consumption and private investment goods are modelled in the same manner.
These final goods are assembled with CES technology, combining intermediate domestic and
imported foreign goods. Varying the use of imported intermediate goods in the production
process is subject to adjustment costs. These final goods are sold taking the price as given.
The public consumption good is a composite of only domestically produced intermediate
goods.
Demand for imported goods is equal to the sum of the respective demands for
intermediate goods for private consumption and investment. These intermediate goods are
sold in the home market by the foreign intermediate-good producer. Domestic and export
15
prices for the same intermediate good might differ as producers use local currency pricing, i.e.
set different prices for the domestic and the export market.
To better understand the structure of the fiscal sector in the NAWM model, it is useful to
the review the government budget constraint:
PG ,t Gt  TRt  Bt  M t 1 
 tC PC ,t Ct   tN (WI ,t N tI  WJ ,t N tJ )   tW (WI ,t N tI  WJ ,t N tJ )   t Wt N t 
Wf
h
(14)
 tK ( RK ,t ut  (  u ( ut )   ) PI ,t ) K t  Tt  Rt1 Bt 1  M t
The left hand side denotes expenditures while the right hand side denotes revenues.
Gt , TR t ,  tC ,  tN ,  tWh and  t
Wf
refer to government consumption, transfers, consumption tax
rate, labor tax rate, employee and employers’ social security contributions and are set
exogenously. Bt and M t are government bonds and money supply. Demands for these assets
are determined by the household’s first order conditions. As in the neoclassical model there
are various assumptions that can be made about what happens to lump sum taxes, Tt .12
Some of the parameters of the model are chosen by setting steady state ratios to match
actual U.S. and Euro area ratios. Others are calibrated along the lines of Smets and Wouters
(2003). The labor supply elasticity equals 0.5 as in the simple neoclassical model of the
preceding section and in the CCTW model. This is a key parameter for the effects of changes
in fiscal policy and we will examine – as in the simulations with the neoclassical growth
model – the sensitivity of the simulation results to variations in the labor supply elasticity.
The share of households without access to financial markets is 25%. While the price
stickiness for goods sold in the domestic market is higher than for exports. The substitution
elasticities between home and foreign goods are set to 1.5. Adjustment costs associated with
changing the import share in investment is a relative high 2.5. With these parameters
consumption and investment respond with low sensitivity to changes in the terms of trade. We
12
In the neoclassical model lump-sum taxes serve to balance the budget. The NAWM model also contains
government debt and lump-sum taxes are related to deviations of the debt to GDP ratio from a target rate. In both
models, the path of lump-sum taxes can be influenced by modifying other distortionary taxes.
16
deviate with respect to the calibration of the intertemporal elasticity of substitution from
Coenen et al (2008). Instead of a value of 2 we use a value of unity which implies log utility
and is consistent with a balanced-growth path. The multiplicative utility function in the
CCTW is already consistent with balanced growth.
Tax rates are calibrated as in Coenen et al (2008). Their values for consumption, income
and social security taxes are based on observable data for the US and the Euro area. The
consumption tax rates is set to 7.7% for the US and 18.3% for the Euro area, the income tax
to 15.4% and 12.2%, respectively, the social security contributions of employees to 7.1% and
11.8%, respectively and the social security contribution of employers to 7.1% and 21.9%
respectively. The capital tax ratio is calibrated to 18.41% in both countries to match the
observed investment-to-output expenditure ratio as in Coenen et al (2008). The government
consumption to GDP ratio is calibrated to 16% for the US and 18% for the Euro area. The
target for the debt-to-GDP ratio is calibrated in both countries to 60% of annual GDP (240%
of quarterly GDP). Transfers in per capita terms are unevenly distributed between households
J and households I in the proportion 3 to 1. Lump-sum taxes in per capita terms are collected
in the proportion of 1 to 3 from households J and households I.
We consider the same illustrative simulation as in the neoclassical model: a reduction
in transfer payments. In the NAWM model, transfers affect the two types of households in
different ways. Figure 6 shows the outcome of a reduction in transfers in the NAWM model.
In this simulation we let the government savings be applied to a reduction in labor taxes. The
reduction in government transfers causes total consumption to rise substantially. Investment
and net exports increase somewhat. Output increases substantially. A fiscal consolidation
results in an increase in total output.
17
Figure 6: A reduction in transfers and labor income taxes
Transfers -1% of GDP, income tax rate -1.6%, phased-in over 5 years
n
W
 + h
Transfers
Consumption
0
1
NAWM
Neoclassical
0.225
0.22
-0.5
0.5
0.215
-1
0
0.21
20
40
0
Investment
20
40
0
0
Output
20
40
Outlays/GDP
0
0
1
-0.5
-0.5
-1
0
0.5
20
40
0
0
-1
20
40
0
20
40
Finally, we consider the CCTW model, which is estimated with U.S. data using
Bayesian methods, as an empirical benchmark for comparison with the calibrated NAWM and
neoclassical model. With regard to short-run effects, it is of interest that the CCTW model
includes about 28 percent of “rule-of-thumb” households which do not smooth consumption.
Because the CCTW model does not incorporate the incentive effects of changes in tax rates
we confine these comparisons to simulations of reductions in government purchases where
the funds saved were paid out in lump sums, such as the simulations in Figure 3 without a
change in the tax rate on capital (solid line).The long-run impact on GDP and government
outlays relative to GDP is very similar in all three models. With regard to the short-run
impact, the CCTW model lies in between the NAWM and neoclassical model. .
3. Fiscal Consolidation Strategy
Having explained the models and illustrated some of their basic properties we can now
consider a budget consolidation reform and estimate its impacts. The Fiscal Consolidation
Strategy is designed to approximate a realistic budget plan that might be employed to reduce
18
federal spending, and thereby, bring the U.S. federal budget deficit down from its current
level of 9 percent of GDP. This prototypical plan contains reductions in both government
purchases and transfer payments from their current trajectory, or baseline. The impact of the
Fiscal Consolidation Strategy is measured against a realistic budget baseline through the year
2022. Thereafter, we assume that the Strategy’s annual expenditure reduction relative to the
baseline remains fixed at its 2022 level.
The 2012-2022 budget expenditure baseline is derived from the Congressional Budget
Office’s (CBO) Alternative Fiscal Scenario.13 This CBO baseline assumes that current
federal expenditure policies will remain in place for the next decade. During this period,
according to CBO, discretionary programs grow at a rate approximated by the rate of
inflation. Mandatory programs, which are driven by three large entitlements; Social Security,
Medicare and Medicaid, grow faster than GDP over the period 2013-2022, mainly as a result
of an increase in the number of workers reaching retirement age and a continuation of rapidly
rising health care costs. Under this baseline, non-interest federal spending gradually declines
by one percent of GDP between 2013 and 2022.
The fiscal consolidation path used in our analysis is modelled on the 2013 Budget
Resolution passed earlier this year by the U.S. House of Representatives.14
The plan calls
for sizeable reductions in expenditures in both discretionary and mandatory program
expenditures relative to the budget baseline between 2013 and 2022. Under the plan, noninterest federal spending is projected to decline to 17.3 of GDP in 2022; a full 2.6 percentage
points of GDP below the CBO baseline in that year.
To estimate the impact of the Consolidated Fiscal Strategy with structural
macroeconomic models, the baseline and House Budget Resolution estimates, both of which
are presented on a unified federal budget basis, must be first converted to a National Income
13
14
Congressional Budget Office, The 2012 Long-Term Budget Outlook, June 2012
House Concurrent Resolution, 112th Congress, 2nd Session, House Report No. 112, March 2012.
19
and Products Account (NIPA) basis. Unfortunately, the publicly available data on the CBO
baseline and the Resolution does not contain sufficient detail to permit a precise conversion to
be made. Nevertheless, a reasonably accurate conversion may be obtained by making a few
simple assumptions.
Our NIPA-based expenditure projections are obtained by converting three large
categories of non-interest federal spending into federal purchases and transfer payments. The
first category consists of four main federal entitlements: Social Security, Medicare, Medicaid,
and health insurance subsidies under the Affordable Care Act. These expenditures, which
account for nearly one-half of all federal spending are treated as transfers.15 The second
category consists of defense expenditures.16 All of the spending in this category, which
accounts for 20 percent of federal spending, is classified as purchases. The third category
consists of all other non-interest expenditures. These expenditures are allocated to purchases
and transfers according to their share in 2011, the latest year for which we have a detailed
conversion of federal budget expenditures into NIPA expenditures.
A final step is to convert the annual budget data series to quarterly data. This step was
accomplished by linearly interpolating the deviations of purchases and transfers from the
budget baseline, measured as a share of GDP, That is, each value in the annual series was
assigned to the last quarterly observation of the associated annual series. Then intermediate
points were placed on straight lines connecting these points.
Purchases and Transfer Payments under the Fiscal Consolidation Strategy
The budgetary impact of the Fiscal Consolidation Strategy, shown earlier in Figure 1,
is substantial. By 2022, federal expenditures relative to GDP are reduced relative to the
15
Medicare expenditures are gross of Medicare receipts, mainly premium payments by enrollees. Under NIPA
accounting, these receipts are recorded as revenues.
16
Defense spending is measured by discretionary spending.
20
baseline by 3.4 percentage points. Figure 7 shows how this reduction is distributed between
federal government purchases and federal government transfer payments.
Figure 7
Fiscal Consolidation Strategy:
Percentage deviation of purchases, transfers and total spending
from baseline as a share of GDP
0.0
Purchases
-0.5
-1.0
-1.5
-2.0
-2.5
Transfers
-3.0
Spending
-3.5
12
14
16
18
20
22
24
26
28
30
32
Year
As the chart shows, the plan’s major budgetary impact is achieved through a reduction in
transfer payments. The plan reduces transfers relative to the budget baseline by 2.5
percentage points. Most of this reduction occurs relatively early in the ten year period; by
2015. The plan’s impact on federal purchases is, in contrast, relatively modest. Government
purchases decline relative to GDP by only .6 percentage points. As was the case with
transfers, most of the decline in purchases occurs early; by 2015.
Debt and Tax Rates under the Fiscal Consolidation Strategy
Because the reform spending path is lower that the baseline spending path, it allows
for lower tax rates and/or lower levels of government debt. We assume a mixture. Under the
Fiscal Consolidation Strategy, the funds released from reduced federal spending are used to
21
reduce the labor tax rate by about 5 percentage points with the remaining funds used to reduce
the debt to GDP ratio as implied by the budget constraint and the model of the economy.
4. Estimating the Impact of the Fiscal Consolidation Strategy
We consider the impacts of the budget reform on the economy by simulating it in the
neoclassical model17 and in the NAWM model. In both cases, we assume that the strategy is
announced and immediately implemented starting in the first quarter of 2013. The
households and firms in the model are assumed to expect immediately that the plan will be
carried out as announced from 2013 onward as illustrated in Figure 1. With the rational
expectations assumption of the model, they take the reduced tax rates and increased after-tax
income into account as they make their consumption decisions. Figure 8 shows the result for
the neoclassical model and Figure 9 shows the result for the NAWM model.
Estimated Impact in the Neoclassical Model
First note that the two charts in the upper left of Figure 8 show the change in
government purchases and transfers and correspond exactly to Figure 7. The chart on the
lower right shows the proposed decline in the tax rate on labor. The neoclassical model
assumes that the government’s budget is balanced every period by lump-sum taxes and does
not incur debt. The timing of labor tax rate changes is chosen to avoid large fluctuations in
lump-sum taxes. We evaluate debt in the NAWM model.
17
For the simulations reported here we use the version of the neoclassical model in which government purchases
appear in the household utility function. In our ongoing research we are examining alternative assumptions, and
testing for robustness of the simulations, which, for this reason, we still consider preliminary
22
Figure 8: Fiscal Consolidation Strategy in the Neoclassical Model
Government purchases
Transfers
0
0
-0.2
-1
Consumption
2
1.5
1
-0.4
-2
0.5
-0.6
2015 2020 2025 2030
-3
Investment
2015 2020 2025 2030
0
Output
2015 2020 2025 2030
Labor Tax Rate
0.2
1.5
0.2
0.15
1
0
0.5
-0.2
0.1
0
2015 2020 2025 2030
2015 2020 2025 2030
0.05
2015 2020 2025 2030
The middle chart in the lower part of the diagram shows that output rises in both the short run
and in the long run. Hence, there is no negative effect of this plan on GDP even though the
budget cuts begin immediately. To the extent that employment moves along with GDP there
is no negative effect on employment either. Because people are forward looking they see that
their incomes will be higher and they begin to increase consumption. The increase in
consumption is more than the decline in government purchases. Moreover people increase the
quantity of their labor supplied as tax rates decline. The permanent impact on the long run is
that GDP rises by two percentage points compared with the baseline assumptions. Note that
consumption also rises in the long run. While there is small decline in investment in the short
run due to a temporary increase in the real interest rate, the overall effect is positive. Thus, the
capital stock rises to a higher long-run level. The initial drop depends on the speed of the
government spending cuts relative to the labor tax reductions. If the tax reduction were to be
23
accelerated, investment would increase right away. Hours worked increase from the start of
the simulation.
Estimated Impact in the Model with Rigidities
Again we can see, now in the upper left of Figure 9, the decline in government
purchases and transfers, and the decline in the tax rate. Note also in the lower left the decline
in the debt as a ratio to GDP.
Despite the wage and price rigidities in this model, there is also an increase in output
in the short run. The reasons are similar to the neoclassical model: the expectations of higher
incomes in the future increase consumption. The long run and the short run effects on output
and consumption are also quite similar in the two models. Thus, our presumption that the
simple neoclassical model would be helpful in pinpointing the long-run impact of expenditure
cuts turns out to be correct in this case. In the models with rigidities the reduction in
government spending also raises permanent income of households, who then decide to
consume more and to enjoy more leisure. The resulting reduction in work hours is causing a
decline in output that is smaller than the decline in government consumption, but still
substantial. Of course, given that the preference parameter on leisure is key in the neoclassical
model and of identical value, lowering it in the DSGE models may similarly reduce the
negative impact of a cut in government purchases on hours worked and total output.
Consumption and output increase on impact with another increase after 5 to 10 years.
Investment decreases temporarily, but also increases in the medium run. The NAWM model
shows that these positive effects are bolstered by an increase in net exports, stimulated by a
real depreciation of the dollar. Of course the effect on net exports in the other “country” in
24
this two country model (i.e. Europe) is the opposite sign, though the overall impacts on euro
area consumption and GDP are small but positive as shown in the appendix.18
Figure 9: Fiscal Consolidation Strategy in the NAWM Model
Government purchases
0
Transfers
0
-0.2
1.5
-1
-0.4
1
-2
-0.6
-0.8
Consumption
2
2015 2020 2025 2030
-3
Investment
0.5
2015 2020 2025 2030
0
Output
Net-Exports
0.4
2
0.2
0.2
1.5
0.15
0
1
0.1
-0.2
0.5
0.05
-0.4
2015 2020 2025 2030
Debt-to-GDP ratio
0
2015 2020 2025 2030
2015 2020 2025 2030
Labor tax rate
0.2
60
0
55
-0.2
50
-0.4
45
-0.6
40
0
Lump-sum tax revenues
65
2015 2020 2025 2030
2015 2020 2025 2030
-0.8
0.15
0.1
2015 2020 2025 2030
0.05
2015 2020 2025 2030
The fluctuations in the lump sum taxes reflect the difficulty of finding a path which
gives both a smooth reduction in the debt to GDP ratio and the tax rate. Fortunately these ups
and downs have little effect because the lump sum taxes are not distortionary and people are
forward looking. In any case the different simulations make clear that the proposed fiscal
consolidation path can be accompanied by sizable tax rate cuts relative to the baseline which
18
For recent evaluations of the impact of government spending cuts in Europe see Roeger and int’Veld (2010)
and Cwik and Wieland (2011). In future work, it would be of interested to investigate the impact of the fiscal
consolidation strategy in the euro area part of the model and quantify the implications of joint U.S. and euro area
consolidation.
25
to increase economic growth in the medium to long run. The anticipation of tax cuts leads also
to growth of consumption and output on impact and avoids any decrease in economic activity.
4. Conclusion
In this paper we estimated the macroeconomic impacts of a fiscal consolidation strategy in
which the government gradually reduces spending over time in order to reduce the deficit and
the growth of the debt. For concreteness the starting point for assessing the strategy is the
current budget situation in the United States where federal spending increased sharply as a
share of GDP during the recent financial crisis and great recession, and spending is expected
to remain high into the future under the baseline projection of the Congressional Budget
Office (CBO).
The fiscal consolidation strategy considered here brings federal spending below this
CBO baseline with spending as a share of GDP returning to the percentage observed prior to
the crisis. While the decrease in spending is gradual, it starts in the first quarter of the new
strategy. Both transfer payments and government purchases of goods and services are reduced
relative to the baseline, though transfers are reduced by a larger amount. The strategy uses the
funds saved both to reduce tax rates and the debt relative to the baseline.
A big question is whether the reduction in government spending reduces GDP in the
short run, a concern that has been raised by many economists and policy makers. We examine
this question and others by simulating the strategy in two types of modern economic models
calibrated to the U.S. economy: (1) models without price and wage rigidities designed to
examine long run issues and (2) and models with such rigidities which are aimed at more
26
explicitly capturing short run effects as well. In both types of models households are forward
looking and they adjust their behavior in response to expectations of future tax and spending
policy.
According to the initial model simulations, the strategy increases GDP in both the
short run and the long run relative to the baseline. There appear to be three sources of this
positive effect. First, lower levels of government spending in the future, compared to the
baseline, imply lower tax rates which provide incentives and stimulate employment. Second,
the expectation of reduced government spending in the future lowers interest rates, which
stimulates demand today offsetting the decline in government spending in the short run. And
third, the lower interest rate reduces the exchange rate thereby increasing net exports which
also offset the decline in government spending. More generally, the gradual and credible
decline in government spending allows the private sector to adjust smoothly to the decline in
spending without negative disruptions.
While the fiscal consolidation strategy focusses on current circumstances in the United
States, the methodology can easily be applied to consolidation plans in other countries or to
joint consolidation plans in several countries.
27
References
Aiyagari, S. Rao, Lawrence J. Christiano, and MartinEichenbaum (1992), “The output,
employment, and interest rate effects of government consumption.”Journal of
Monetary Economics, 30:73-86.
Amano, Robert A. and Tony S. Wirjanto (1998).“Government expenditures and the
permanent-income model.”Review of Economic Dynamics, 1: 719-730.
Barro, Robert J. (1990). “Government spending in a simplemodel of endogeneous
growth.”Journal of Political Economy, 98(S5): 103-125.
Baxter, Marianne and Robert G. King (1993), “Fiscal Policy in General Equilibrium”, The
American Economic Review, 83(3): 315-334.
Chetty, Ray, Adam Guren, Day Manoli and Andrea Weber (2011a), “Are micro and macro
labor supply elasticities consistent? A review of evidence on the intensive and
extensive margins”, The American Economic Review Papers and Proceedings, 101(3):
471-475.
Chetty, Ray, Adam Guren, Day Manoli and Andrea Weber (2011b), “Does indivisible labor
explain the difference between micro and macro elasticities?”, NBER Working Paper
No. 16729.
Christiano, Lawrence J., Martin Eichenbaum, and Charles L. Evans (2005), “Nominal
Rigidities and the Dynamic Effects of a Shock to Monetary Policy,” Journal of
Political Economy, 113(1): 1-45.
Coenen, Günter, Peter McAdam and Roland Straub (2008), “Tax reform and labour-market
performance in the euro area: A simulation-based analysis using the New Area-Wide
Model”, Journal of Economic Dynamics and Control, 32: 2543-2583.
Coenen, Guenter and Roland Straub (2005), “Does Government Spending Crowd in Private
Consumption? Theory and Empirical Evidence for the Euro Area,” International
Finance, 8(3): 435-470.
Cogan, John, Tobias Cwik, John B. Taylor, and Volker Wieland (2010), New Keynesian
versus Old Keynesian Government Spending Multipliers, Journal of Economic
Dynamics and Control, March 2010, 34: 281-295.
Cwik, Tobias and Volker Wieland, (2011), Keynesian Government Spending Multipliers and
Spillovers in the Euro Area, Economic Policy, July 2011, 26: 493-549.
Ljungqvist, Lars and Thomas J. Sargent (2004).Recursive Macroeconomic Theory, MIT
Press, 2nd edition.
Hall, Robert E. (1980), “Labor supply and aggregate fluctuations,” In On the State of
Macroeconomics, Carnegie Rochester Conference on Public Policy, XII, Amsterdam,
North Holland.
King, Robert G., Plosser, Charles I. and Rebelo, Sergio T. (1988).“Production, growth and
business cycles : I. The basic neoclassical model,”Journal of Monetary Economics,
21(2-3): 195-232.
Linnemann, Ludger and Schabert, Andreas (2004). “Can fiscal spending stimulate private
consumption?,”Economics Letters, 82(2): 173-179.
Roeger W., in 't Veld, J. (2010). “Fiscal stimulus and exit strategies in the EU: a model-based
analysis", European Economy Economic Papers no. 426.
Smets, Frank and RafWouters (2007), "Shocks and Frictions in U.S. Business Cycles: A
Bayesian DSGE Approach." American Economic Review 97, 3: 506–606.
Taylor, John B. (1993), Macroeconomic Policy in a World Economy: From Econometric
Design to Practical Operation, WW Norton, New York.
28
Taylor, John B. (2009), “The Lack of an Empirical Rationale for a Revival of Discretionary
Fiscal Policy,” American Economic Review, Papers and Proceedings, (May) Vol. 99,
No. 2.
Taylor, John B. and Volker Wieland (forthcoming), "Surprising Comparative Properties of
Monetary Models," The Review of Economics and Statistics.
Uhlig, Harald and Thorsten Drautzburg (2011), “Fiscal Stimulus and Distortionary Taxation,
NBER Working Paper 17111.
Wieland, Volker, Tobias Cwik, Gernot Mueller, Sebastian Schmidt and MaikWolters
(forthcoming), "A New Comparative Approach to Macroeconomic Modelling and
Policy Analysis" Journal of Economic Behavior and Organisation.
29
Appendix
Figure A1: Responses to a joint reduction of government spending and the labor tax rate.
Government spending
Labor Tax Rate
0
Hours
0.16
0.2
0%
1%
2%
-0.5
0.15
0
-0.2
0.14
-0.4
-1
0
20
40
0.13
0
20
Consumption
40
-0.6
0
20
Investment
1
40
Output
1
1
0.5
0.5
0
0
-0.5
-0.5
0.5
0
0
20
40
-1
0
20
40
-1
0
20
40
Figure A2: Tax revenues in the case of a labor tax rate reduction.
Lump-sum tax revenue
Capital tax revenue
0
0
-0.5
-1
-1.5
0
10
20
30
40
Labor tax revenue
0%
1%
2%
Consumption tax revenue
0
0
-0.5
-0.5
-0.5
-1
-1
-1
-1.5
0
10
20
30
-1.5
40
0
10
20
30
40
-1.5
0
10
20
30
Figure A3: Tax revenues in the case of different labor supply elasticities and a joint
government spending and capital tax reduction.
Lump-sum tax revenue
Capital tax revenue
elasticity=0.5
elasticity=0.25
totally inelastic
0.2
Labor tax revenue
Consumption tax revenue
0.2
0.2
0
0
0
-0.2
-0.2
-0.2
-0.2
-0.4
-0.4
-0.4
-0.4
-0.6
-0.6
-0.6
-0.6
-0.8
-0.8
-0.8
-0.8
-1
-1
-1
0
0
10
20
30
40
0
10
20
30
40
0.2
0
30
10
20
30
40
-1
0
10
20
30
40
40
Figure A4: Tax revenues and utility of public consumption
Lump-sum tax revenue
Capital tax revenue
alpha=1
alpha=0.75
0.2
Labor tax revenue
Consumption tax revenue
0.2
0.2
0
0
0
0
-0.2
-0.2
-0.2
-0.2
-0.4
-0.4
-0.4
-0.4
-0.6
-0.6
-0.6
-0.6
-0.8
-0.8
-0.8
-0.8
-1
0
10
20
30
-1
40
0
10
20
30
40
-1
0.2
0
10
20
30
40
-1
0
10
20
30
30
40
30
40
40
Figure A5: The role of substitutability of private and public consumption
Capital Tax Rate
Government spending
Capital Stock
0.19
0
gamma =
gamma =
gamma =
gamma =
gamma =
-0.2
-0.4
-0.6
1
0.75
0.36
-0.41
-1
5
4
0.18
3
2
0.17
1
-0.8
-1
0
0.16
0
10
20
30
40
0
10
Consumption
20
30
40
0
Investment
1.5
1.5
1
1
1
0.5
0.5
0.5
0
0
10
20
30
40
20
Output
1.5
0
10
0
0
10
20
30
40
0
10
20
Substitutability: The effects of government consumption in the utility function depend on the
degree of substitutability of private and public consumption as we will show in the following
analysis. Intuitively, if private and public consumption are substitutes, a decrease in public
consumption will lead to an increase in private consumption. If they are complements, a
decrease in public consumption leads to a decrease in private consumption. Figure A5 shows
this. We calibrate the weight of private to public consumption in the utility function to  =0.8
and vary the degree of substitutability. We include the value estimated by Amano and
Wirjanto (1998) and the mean of the estimates by Ni (1995) of   0.41 . Estimates of other
authors use different and less general functional forms of the utility function and are, thus, not
comparable. We include, however, also other values to cover a large range of substitutability
degrees. The results show the non-monotonic effect of decreasing of  , i.e. the substitutability
of private and public consumption as described above. For highly negative values of  one
can even generate a comovement of government spending and consumption (not shown).
These values are, however, totally off the estimated value of   0.36 by Amano and
Wirjanto (1998).19,20
19
Gamma=0.36 is consistent with an estimate of the intratemporal elasticity of substitution between private and
public consumption of about 1.56.
31
Figure A5: Impact of U.S. Fiscal Consolidation Strategy on the Euro Area
EA Consumption
EA Investment
0.5
EA Output
0.5
0.3
EA Net-Exports
0.1
0
0.2
0
0
-0.1
0.1
-0.2
0
-0.1
20
2015 2020 2025 2030
-0.5
-0.5
2015 2020 2025 2030
2015 2020 2025 2030
2015 2020 2025 2030
Note that in our setting 1 /  1  0.5    1 /(1   )  1.56 and thus private and public consumption are
substitutes. Amano and Wirjanto (1998) estimate 1 /  1  1.56    1 /(1   )  1.56 , so that private and
public consumption are unrelated. However, their estimate
value of 2.
32
 1 is very low, so that we proceed with the usual